Using Vox Proxy with Windows Vista

Microsoft has made several changes to the Microsoft Agent technology used in Vox Proxy. These changes primarily affect the speech techlology supported, and represent some good news and some bad news:

First the good news:

The speech API (the programming interface) for Agent has been upgraded to support both SAPI4 andSAPI5. The XP version of Microsoft Agent supports only SAPI4 voices. This change will permit the use of several newer Text-To-Speech (TTS) engines and voices from vendors whose voices were not previously supported by Vox Proxy. It also permits the use of new speech "tags" for changing pitch, speed, and other characteristics of the voices.

Beginning with Version 4.320, Vox Proxy contains enhancements to detect SAPI5 speech engines and allow you to use them in your scripts. The "Show a new character" wizard detects and displays all SAPI4 and SAPI5 voices in the "Voices" drop-down list. The TTSEngine command wizard (in "Other Agent Commands" also displays both SAPI4 and SAPI5 voices. There are many more SAPI5 Text-To-Speech engines currently available from third-party vendors than SAPI4, so you may find this a useful new capability. Note that AT&T Natural Voices, available on the Vox Proxy web site, automatically install in both SAPI4 and SAPI5 versions, so these are not an issue running under Vista.

download the latest free program update here (Current version 4 users only - do not download this if you are on versions 1, 2, or 3) Save this file in the folder: c:\program files\vox proxy and replace the old file of the same name.

Now the bad news:

SAPI5 voices are not backward-compatible. That is, scripts you develop in Vista using a SAPI5 speech engine will not speak when played on a Windows XP machine, whether in the full Vox Proxy software or the Vox Proxy Player. This is a Microsoft change for Vista about which we have no control. You can, of course, create WAV files using various third-party software from the SAPI5 speech engines and speak using those wav files.

Another bit of bad news is that Microsoft has removed the multi-lingual features in Agent characters. With Vista, the characters will speak only in the language set as the default for the computer on which it is running. So if your computer is set for US English, the character will speak only with US English Text-To-Speech voices. If you use AT&T Natural Voices with Vox Proxy, there is a work-around for this problem: you can use Vox Proxy's conversion of speech to WAV files and let the characters speak using the WAV files. With other SAPI5 Text-To-Speech engines, you can similarly create WAV files using third-party software. Though less convenient than speaking directly from the TTS engine, there are other benefits to speaking with WAV files, and it at least provides a way to speak in multiple languages.

You can read more about using Microsoft Agent with Vista directly from Microsoft by clicking here.

 
Right Seat Software, Inc.
Information updated 1/08/2009